Post on 07-Jun-2020
Thermoelectric properties of indium doped PbTe1-ySey alloysAshoka Bali, Heng Wang, G. Jeffrey Snyder, and Ramesh Chandra Mallik
Citation: Journal of Applied Physics 116, 033707 (2014); doi: 10.1063/1.4890320 View online: http://dx.doi.org/10.1063/1.4890320 View Table of Contents: http://scitation.aip.org/content/aip/journal/jap/116/3?ver=pdfcov Published by the AIP Publishing Articles you may be interested in Importance of ligands in the electronic properties of FeTe0.6Se0.4 J. Appl. Phys. 114, 163906 (2013); 10.1063/1.4827192 Correlation between microstructure and thermoelectric properties of AgPb18SbTe20 (LAST-18) AIP Conf. Proc. 1449, 171 (2012); 10.1063/1.4731524 High-output-power densities from molecular beam epitaxy grown n- and p-type PbTeSe-based thermoelectricsvia improved contact metallization J. Appl. Phys. 111, 104501 (2012); 10.1063/1.4712425 The QuantumFunctional Properties of Pr1xyLaxPbyTe AIP Conf. Proc. 850, 1291 (2006); 10.1063/1.2355179 Macroscopic thermoelectric inhomogeneities in ( Ag Sb Te 2 ) x ( Pb Te ) 1 x Appl. Phys. Lett. 87, 171903 (2005); 10.1063/1.2056590
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
Thermoelectric properties of indium doped PbTe1-ySey alloys
Ashoka Bali,1 Heng Wang,2 G. Jeffrey Snyder,2 and Ramesh Chandra Mallik1,a)
1Thermoelectric Materials and Devices Laboratory, Department of Physics, Indian Institute of Science,Bangalore 560012, Karnataka, India2Department of Materials Science, California Institute of Technology, Pasadena, California 91125, USA
(Received 31 May 2014; accepted 3 July 2014; published online 18 July 2014)
Lead telluride and its alloys are well known for their thermoelectric applications. Here, a
systematic study of PbTe1-ySey alloys doped with indium has been done. The powder X-Ray dif-
fraction combined with Rietveld analysis confirmed the polycrystalline single phase nature of the
samples, while microstructural analysis with scanning electron microscope results showed densifi-
cation of samples and presence of micrometer sized particles. The temperature dependent transport
properties showed that in these alloys, indium neither pinned the Fermi level as it does in PbTe,
nor acted as a resonant dopant as in SnTe. At high temperatures, bipolar effect was observed which
restricted the zT to 0.66 at 800 K for the sample with 30% Se content. VC 2014 AIP Publishing LLC.
[http://dx.doi.org/10.1063/1.4890320]
I. INTRODUCTION
The need for alternative sources of energy has led to an
interest in thermoelectric materials, where direct conversion
of heat to electricity, and vice-versa is possible. The effi-
ciency of this conversion depends on a dimensionless figure
of merit (zT)¼S2T/qjtotal, where S is the Seebeck coeffi-
cient, q is the electrical resistivity, T the temperature and
jtotal (¼ jeþ jl), is the total thermal conductivity of the ma-
terial, with je as the electronic and jl the lattice contribution.
Lead telluride (PbTe) and its compounds have been used as
mid temperature (300–850 K) thermoelectric materials with
a zT of �0.8.1 The material parameter (b), defined by
b ¼ kB
e
� �22elð Þjl
2pm�kBT
h2
� �3=2
T; (1)
where l is the mobility, kB the Boltzmann constant, e the
electronic charge, jl the lattice thermal conductivity, m* the
density of states effective mass given by Nv2=3ðm�2
?m�jjÞ1=3
,
Nv is the band degeneracy and m*? and m*jj are effective
masses in the transverse and the longitudinal directions,
respectively, is related to zT. At any temperature, depending
on the scattering mechanism, all the properties except jl are
electronic properties, which depend on the carrier concentra-
tion. b can be increased by reducing jl by alloying it with an
element from the same group at either site of the PbTe crys-
tal—Pb or Te. Secondly alloying also results in a change of
band structure like tuning of band gap (e.g. in PbSnxTe1-x
alloys2), convergence of bands,3 etc. PbTe-PbSe alloys,
where Te is substituted by isovalent Se, have been studied
several times in the past.4–8 A band degeneracy of 12 was
reached by convergence of many valleys in valence band at
high temperatures in p-type PbTe1-ySey material resulting
in high zT of 1.8 at 850 K.3 In another report, zT¼ 1.35 at
675 K (Ref. 6) was obtained by adding 0.7 at. % Na. On the
other hand, n-type alloys have shown lower zT,9 primarily
due to the absence of any such band modification. Moreover,
in both the cases, the high value of zT is obtained only at
high temperatures, whereas for practical applications, high
zT in the entire operation range is desirable. Hence, for
achieving better n-type PbTe1-ySey alloys, additional doping
is required to be done in order to control the above men-
tioned electronic properties. Even though there are several
well established n-type dopants in PbTe,10,11 indium (In) has
been found to lead to a stabilization of carrier concentration
up to 2 at. % In content, which happens due to Fermi level
pinning.12,13 This could also be useful for PbTe1-ySey alloys
for achieving homogeneity of transport properties over a
larger temperature range. Second, even though other group
III elements like Tl have been found to form resonant level
in PbTe,14 In has not been found to do so in either PbTe or
PbSe. However, since alloying leads to a change of band
gap and also the band structure, there is a possibility of
doing so, which could lead to increase in Seebeck coeffi-
cient. Further, the effect of substituting In at Pb site and Se
at Te site simultaneously on the band parameters and their
consequent effect on thermoelectric properties have not
been studied so far.
In this paper, a systematic study of PbTe-PbSe system
with varying Se and fixed In content has been done. PbTe
and PbSe have been known to form solid solution over a
large composition range.5 Here, a maximum of 30% Se was
chosen for the PbTe-PbSe alloy system. For indium, its
phase diagram shows that InTe is soluble in PbTe upto
7 mol. % (Ref. 15) with single phase structure. Here, a low
concentration of 0.1 at. % In was chosen, which falls well
within this region. No supplementary doping with halogens
was done so as to study the control of carrier concentration
only by indium and the off-stoichiometry from PbTe. It was
found that the carrier concentration of the samples varied in
a narrow range since Se did not contribute to any extra elec-
trons to the system and a maximum zT of 0.66 was reached
at 800 K.
a)Author to whom correspondence should be addressed. Electronic mail:
rcmallik@physics.iisc.ernet.in
0021-8979/2014/116(3)/033707/7/$30.00 VC 2014 AIP Publishing LLC116, 033707-1
JOURNAL OF APPLIED PHYSICS 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
II. EXPERIMENTAL DETAILS
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25,
and 0.30, labeled as VA1, VA2, VA3, VA4, VA5, and VA6,
respectively) was prepared from high purity Pb(99.9%,
Te(99.999%), In(99.99%), and Se(99.999%). The elements
were mixed in appropriate stoichiometric ratios and sealed
under 1.3� 10�2 Pa vacuum in carbon coated quartz tubes.
The ampoules were then slowly heated upto 1273 K in 10 h
and soaked at that temperature for 6 h, which was followed
by quenching in water. The samples were then annealed at
973 K for 72 h, after which water quenching was done again.
The ingots thus prepared were subjected to hand grinding
with mortar and pestle and compacted into cylindrical pellets
by hot pressing under vacuum �7� 10�5 torr at 913 K with
30 MPa pressure for 30 min. Two identical pellets were pre-
pared for each sample—one was used for transport properties
measurement, while the other was used for microstructural
characterisation. The pellets used for transport property mea-
surement were 14 mm diameter and thinned down to 1 mm
thickness. All the pellets had more than 95% of the theoreti-
cal density (8.24 g/cm3).
Phase identification was done by Bruker D8 Advance
with Cu Ka radiation with Si as an internal standard and scan
speed of 2�/min. Backscattered electron images (BSEs) were
taken with Quanta 200 environmental scanning electron
microscope (ESEM). Compositional analysis was carried out
on JEOL JXA-8530F Electron Probe Micro Analyzer
(EPMA) with Wavelength Dispersive Spectroscopy (WDS)
detectors. Temperature dependent Seebeck coefficient was
measured using the apparatus described in Ref. 16. The in-
plane Hall coefficient and electrical resistivity were meas-
ured using the Van der Pauw method under 62 T magnetic
field. Thermal diffusivity (D) was measured by laser flash
method (Netzsch LFA 457) on the same pellets. Prior to
carrying out the measurement, the pellets were coated with
graphite spray for better heat conductance. The specific heat
(Cp) was calculated using the empirical formula
Cp=kB ¼ 3:07þ 4:7� 10�4ðT=K� 300Þ, which is based on
experimental results and has been used earlier for lead chal-
cogenides.3,7,10,11,17 The thermal conductivity was calculated
from j¼Cp� d�D, where d is the density. For each of the
transport property measurement, the uncertainty was 5%.
III. RESULTS AND DISCUSSION
A. Material characterisation
Figure 1(a) shows the X-ray diffraction pattern of all the
samples. Sharp peaks indicative of polycrystalline nature of
the samples were observed. The peaks were indexed accord-
ing to ICSD #182660 for PbTe since both the binary end
compounds—PbTe and PbSe—have NaCl type structures
(space group Fm-3m) and similar lattice constants
(PbTe¼ 6.460 A, PbSe¼ 6.217 A) and hence, almost similar
XRD patterns. Rietveld refinement using the software
FULLPROF18 was done to obtain the lattice constants for all
the samples. Figure 1(b) shows the refinement result for sam-
ple VA2. The variation of lattice constant with Se content
for all the samples is shown in the inset. All the samples
followed Vegard’s law, which indicates solid solution forma-
tion between PbTe and PbSe. The decrease in lattice constant
occurs due to smaller radii of both the ions which substitute
at Pb and Te sites. In3þ (ionic radius¼ 94 pm) replaces a
larger Pb2þ (ionic radius¼ 133 pm) ion, while a smaller
Se2� (ionic radius¼ 184 pm) replaces the larger Te2� (ionic
radius¼ 207 pm) ion—both of which are expected to give
rise to shrinkage of lattice.
Figure 2 shows scanning electron micrographs of repre-
sentative sample VA2 (Pb0.999In0.001Te0.95Se0.05). Figure
2(a) corresponds to the as-grown sample showing small
pores in several places. Figure 2(b) shows their fracture sur-
face, indicating the average particle size to be in micro-
meters. Figure 2(c) shows the polished surface of the hot
press sample indicating single phase and no distinct grain
boundaries. Figure 2(d) is the micrograph of fracture surface
of the hot press pellet, which showed reduction of the pores
and densification during the hot press. The nominal and the
observed composition for all the samples taken at several
points and averaged are listed in Table I. In all the samples,
even though within error limits, a slight excess of Pb was
observed. The observed amount of selenium and tellurium
matched well with the nominal composition, while indium
was found to be lower. The loss of indium due to high vapor
pressure (v.p.) has been ruled out by comparing with v.p. of
other elements at high temperatures. At 1233 K, indium
FIG. 1. (a) X-ray diffraction patterns of the samples Pb0.999In0.001Te1-ySey
(y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30). (b) X-ray diffraction pattern of
sample VA2 (Pb0.999In0.001Te0.95Se0.05) after Rietveld refinement. The inset
shows the variation of the lattice constant with Se composition for all the
samples.
033707-2 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
achieved a vapor pressure of 1.35� 10�2 torr,19 while Te
and PbTe measure at �0.63 torr and �1� 10�2 torr, respec-
tively, at 1215 K.20 Since Te has a higher v.p. than indium
and could thus vaporize more easily during synthesis, but
was found to be present in almost similar values as the nomi-
nal composition, it is less probable that In would be lost in
the synthesis process. Instead, the less amount of In observed
here could be because it had been added in extremely small
amount (0.1 at. %), and it is difficult to obtain very accurate
values for the same from using internal standards during
EPMA measurement.
B. Transport properties
Figure 3 shows temperature dependent Seebeck coeffi-
cient from 300–800 K. All the samples were of n-type indi-
cating electrons as the majority carriers throughout the
temperature range. The values from literature for several
related PbTe and PbSe systems8,10,21,22 are plotted alongside
the experimental data. The Seebeck coefficient obtained in
this study resembles more closely to reports on PbSe system
and iodine doped n-type PbTe, which is a better n-type donor
as compared to indium in PbTe. This is because of similar
carrier concentration (4� 1019 cm�3 to 5� 1019 cm�3)
obtained in this study. For In doped PbTe, (Gelbstein et al.21
and Guch et al.22) higher Seebeck coefficient was reported
throughout the temperature range corresponding to larger
room temperature values (��150 to �200 lV/K) due to
their lower carrier concentration (�1018 cm�3) as compared
to n� 1019 cm�3 reported for the halogen doped PbTe10 and
PbSe.9 The Seebeck values for all the samples do not show
any systematic variation with Se content since the carrier
concentration does not changes systematically, as seen later
in the results obtained by Hall measurement. Figure 4 shows
the carrier concentration (n) from 300–800 K. The samples
do not show much change in n on variation of Se content
since the indium content was kept constant in all the samples
and unlike halogens, Se belongs to the same group as Te so
that no extra electrons are given to the PbTe1-ySey system.
The observed minor differences might be due to defects, dis-
locations etc., occurring during synthesis which cannot be
quantified easily and also due to a slight Te/Se loss resulting
in off-stoichiometry. Since the solubility of Pb is very less in
PbTe, even a slight excess of Pb might give excess of elec-
trons, which would have resulted in the higher carrier con-
centration observed here than that expected from only In
doping. The temperature dependence of n is directly related
to the temperature dependence of Seebeck coefficient. This
is because n remains constant till 700 K, which agrees with
linear increase in Seebeck. The rise in n values above 700 K
is responsible for the apparent flattening of Seebeck values.
Figure 5 shows the temperature dependent resistivity
from 300–800 K. An increase in magnitude with increasing
temperature shows the highly doped nature of the samples.
The resistivity values obtained here are comparable to those
FIG. 2. Scanning electron micrographs of representative sample VA2
(Pb0.999In0.001Te0.95Se0.05) showing (a) as-grown sample (b) fracture surface
of as-grown sample (c) polished surface of hot pressed pellet and (d) fracture
surface of hot pressed pellet.
TABLE I. Results from compositional analysis, fitting of log l-log T curves and the respective carrier concentration at room temperature.
S. No. Sample Nominal composition EPMA composition
Slope from log l-logT fit
n (at 300 K) (1019 cm�3)Region I (low T) Region II (high T)
1. VA1 Pb0.999In0.001Te0.99Se0.01 Pb1.149In0.0001Te1.002Se0.007 �0.96 �2.64 4.353
2. VA2 Pb0.999In0.001Te0.95Se0.05 Pb1.147In0.002Te0.947Se0.049 �0.85 �2.51 3.911
3. VA3 Pb0.999In0.001Te0.90Se0.10 Pb1.096In0.0008Te0.896Se0.095 �0.78 �3.02 5.006
4. VA4 Pb0.999In0.001Te0.80Se0.20 Pb1.129In0.0008Te0.800Se0.182 �0.97 �1.72 4.178
5. VA5 Pb0.999In0.001Te0.75Se0.25 Pb1.098In0.0002Te0.737Se0.288 �0.59 �2.63 3.901
6. VA6 Pb0.999In0.001Te0.70Se0.30 Pb1.0375In0.0005Te0.699Se0.267 �1.11 �2.46 4.401
FIG. 3. Temperature dependent Seebeck coefficient for the samples
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30). The litera-
ture values are plotted alongside (open symbols) for comparison.
033707-3 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
obtained by Wang et al.9 for a similar carrier concentration
in Br doped PbTe-PbSe alloys and lower than that of
Gelbstein et al.21 for 0.1 at. % In doped PbTe. To some
extent, with increase of selenium content, the resistivity
increases, in contrast to Seebeck results.
Figure 6 shows Pisarenko plots (S–nH) at different tem-
peratures. PbTe and PbSe are narrow band gap materials
with a direct band gap at L point of the Brillouin zone and
the importance of nonparabolicity of bands is important for
their alloys too.23 Therefore, the theoretical curves here are
obtained from single Kane band (SKB) model described in
Refs. 9 and 24, assuming conduction by a single band for
which the transport coefficients are given as follows:
The Seebeck coefficient is given by
S ¼ kB
e
1F1�2
0F1�2
� n
" #; (2)
where e is the electronic charge, n is the reduced chemical
potential, 1F1�2 is the integral of the form
nFm1 ¼
ð10
� @f
@e
� �en eþ ae2ð Þm 1þ 2aeð Þ2 þ 2
h i1=2
de; (3)
where f is the Fermi function, e is the carrier energy and
a¼ kBT/Eg. For lead chalcogenides, the band gap (Eg) is
temperature dependent, given by
eg=eV ¼ 0:17þ 4� 10�4T=K: (4)
Carrier density is given by
n ¼ 2m�kBTð Þ3=2
3p2�h30F
3=20 : (5)
and mobility by
l ¼ 2p�h4eC1
m�I 2m�bkBTð Þ3=2N2
30F1�2
0F3=20
; (6)
where �h is the reduced Planck’s constant, Cl the average lon-
gitudinal elastic constant, m�I ¼ 3 2m�?þ 1
m�k
� ��1is the inertial
effective mass, N is the deformation potential and
m� ¼ N2=3V m�b. The values for all the constants have been
taken from Ref. 9. The values of n and l were multiplied by
Hall factor (A), defined by
A ¼ 3K Kþ 2ð Þ2Kþ 1ð Þ2
0F1=2�4
0F3=20
0F1�2
� �2; (7)
where the factor K is the ratio of effective masses
(K¼m�jj=m�? ¼ 1.75), assuming T independence and3KðKþ2Þð2Kþ1Þ2
is the anisotropy factor. This is because the Hall factor
accounts for the nonspherical shape of the constant energy
surfaces of lead chalcogenides as well as the dependence of
relaxation time on energy24 such that its value deviates
from 1. Figure 7 shows the dependence of A on g at different
temperatures. The increase of these values for higher T and
high g values shows that the usual assumption of A¼ 1 is
not valid in this system too, so that the Hall carrier concen-
tration (nH¼ n/A) was used for all the calculations. At lower
temperatures, the Pisarenko plots describe the data well,
which indicated that In does not shows any resonant
FIG. 5. Temperature dependent electrical resistivity for the samples
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30). The litera-
ture values are plotted alongside (open symbols) for comparison.
FIG. 6. The plots of S-nH (Pisarenko plots) for different temperatures. The
open symbols correspond to the experimental data for all the samples
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30) at each
temperature, described as follows: �—300 K, w—400 K, �—500 K, �—
600 K, and �—700 K. The curves have been calculated theoretically.
FIG. 4. Temperature dependent carrier concentration for the samples
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30).
033707-4 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
behavior in Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20,
0.25, and 0.30) system, unlike other reports of In doping on
SnTe,25 PbSnxTe1-x (x¼ 0.18), etc. This might be because
the band structure of PbTe1-ySey alloy is not much different
from PbTe and PbSe. The ln(m*) vs ln(T) curve is given in
Fig. 8. The slope is 0.56 which agrees with the theoreticaldðlnm�ÞdðlnTÞ ¼ 0:5 dependence for lead chalcogenides. This strong
temperature dependence of m* on T occurs due to both inter-
actions between the conduction and valence band, thermal
expansion of the lattice and interaction between carriers and
phonons.24
Figure 9 shows the temperature dependent mobility for
all the samples. All the samples showed power law decrease
l ¼ T� with temperature, which arises because of the energy
and temperature dependence of relaxation time (s) and m*,
respectively. The coefficient � showed two different set of
values (Table I) showing scattering by acoustic phonons, as
expected at high temperatures. The �� 1 can be accounted
for by a mixed scattering mechanism of acoustic phonon and
ionized impurity scattering. At higher temperatures, all the
lead chalcogenides show T�2.5 behavior for acoustic
scattering due m�ð�2Þ temperature dependence in the relation
l / m�ð�2ÞT�1 for degenerate semiconductors.24 The mobil-
ity values at all temperatures in this study are lower than
those observed in n-type PbSe,7,9,26 PbTe10 first, because In
does not completely dope in PbTe1-ySey as compared to
other n-type dopants like halogens (Br, I) or p-type dopants
like Na, K etc., and second, due to alloy scattering. With
increase in temperature, decrease in mobility is observed in
all samples.
Figure 10(a) shows the temperature dependent total ther-
mal conductivity for all the samples. The electronic part was
calculated using the Wiedemann-Franz relation je¼LT/q,
where L is the Lorenz number. The lattice contribution
(Fig. 10(b)) was then obtained by subtracting it from jtotal.
Since the electronic contribution is affected by the nonpara-
bolicity of the bands, Lorenz number was calculated from
SKB model assuming acoustic phonon scattering using the
equation,9,24
FIG. 7. Hall factor as a function of reduced Fermi energy at different
temperatures.
FIG. 8. Plot of ln (m*) – ln(T). Each point in the graph corresponds to the
effective mass for the Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25,
and 0.30) series of samples calculated from the S-nH curves at different
temperatures.
FIG. 9. Temperature dependent mobility for the samples Pb0.999In0.001Te1-ySey
(y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30).
FIG. 10. (a) Temperature dependent total thermal conductivity for the sam-
ples Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30). (b)
Temperature dependent lattice thermal conductivity for the samples
Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30). The curves
indicate the theoretically calculated curves.
033707-5 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
L ¼ kB
e
� �2 2F1�2
0F1�2
�1F1�2
0F1�2
!224
35; (8)
where the parameters have already been described. At 300 K,
jl values varied between 4.7 and 3.4 W/mK, which is slightly
higher as compared to other reports.9–11 At high tempera-
tures, jl values reduce to �1 W/m-K, which is still high as
compared to the theoretically calculated minimum jl of
0.38 W/m-K possible for PbTe at its melting point,27 consid-
ering both the longitudinal and optical branches. This shows
that there is a further scope in such compounds to reduce the
lattice thermal conductivity, possibly by reducing grain size,
nanostructuring, etc., to achieve higher zT. Microstructural
engineering approach has been previously used28–33 for sev-
eral thermoelectric materials. The expected decrease in jl
with temperature due to Umklapp scattering was observed
upto �700 K. The possible role of indium doping here could
be to scatter phonons to reduce thermal conductivity.
A simple model by Leibfried and Schl€omann for jl de-
pendence at high temperature is given by34
j1 ¼ 3:5kB
h
� �3MV1=3H3
D
c2T; (9)
where M is the average mass per atom, V the average vol-
ume per atom, c the Gruneisen parameter, which shows the
anharmonic nature of the crystal and HD is the Debye tem-
perature. Here, the volume was taken from the Rietveld
refinement for sample VA2, while HD¼ 136 K (Ref. 35) and
c¼ 1.45 for PbTe since the content of Te is higher at this
composition. Since this model does not explicitly consider
different scattering mechanisms, it shows only 1/T decrease,
but may not agree accurately if any specific scattering mech-
anism dominates. Second, in PbTe1-ySey solid solutions,
with increase of Se content, jl first decreases, comes to a
minimum at �y¼ 0.5, and then rises again26,36 due to high-
est disorder at mid compositions. Here, electron-phonon
interaction can be neglected due to low values of je while
scattering by acoustic and also by optical polar phonons are
considered.
Figure 11 shows the thermoelectric figure of merit (zT)
for all the samples. The highest zT of 0.66 was achieved at
800 K for sample with 30% Se content due to optimization
of the carrier concentration for these samples. This is higher
than 0.45 at 650 K obtained by Guch et al. for In doped PbTe
but lower than 0.92 at 575 K by Gelbstein and �1.0 at 850 K
by Wang et al. for Br doped PbSe. A possible reason is the
occurrence of bipolar effect at high temperatures in the pres-
ent set of samples, which was absent in all the other reports.
IV. CONCLUSIONS
A number of n-type indium doped PbTe1-ySey alloys
were synthesized by solid state synthesis method and the
thermoelectric properties were studied. Indium was not
found to lead to Fermi level pinning in PbTe1-ySey as previ-
ously reported in PbTe. Indium is also not a resonant scat-
terer in these alloys. Bipolar conduction was found at high
temperature which was not observed in PbTe or PbSe doped
with halogens (good dopant) for n-type or Na doping for
p-type materials. This indicates that it is very important to
control the carrier concentration in the alloys for good ther-
moelectric properties, particularly at high temperatures.
Future work could involve proper control of carrier concen-
tration to avoid bipolar contribution at high temperature, and
a further reduction of lattice thermal conductivity by other
approaches like nanostructuring, which could increase the zT
further.
ACKNOWLEDGMENTS
One of the authors (AB) would like to acknowledge Indo-
US Science and Technology Forum (IUSSTF) and American
Physical Society (APS) for the grant received under the
“APS-IUSSTF Physics Student Visitation Program.”
1Z. H. Dughaish, Physica B 322, 205 (2002).2V. Jovovic, S. J. Thiagarajan, J. P. Heremans,T. Komissarova, D.
Khokhlov, and A. Nicorici, J. Appl. Phys 103, 053710 (2008).3Y. Pei, X. Shi, A. LaLonde, H. Wang, L. Chen, and G. J. Snyder, Nature
473, 66 (2011).4R. N. Tauber and I. B. Cadoff, J. Appl. Phys 38, 3714 (1967).5J. Steinnger, Metall. Trans. 1, 2939 (1970).6I. Kudman, J. Mater. Sci. 7, 1027 (1972).7H. Wang, Y. Pei, A. D. LaLonde, and G. J. Snyder, Adv. Mater. 23, 1366
(2011).8Q. Zhang, F. Cao, W. Liu, K. Lukas, B. Yu, S. Chen, C. Opeil, D. Broido,
G. Chen, and Z. Ren, J. Am. Chem. Soc. 134, 10031 (2012).9H. Wang, Y. Pei, A. D. LaLonde, and G. J. Snyder, Proc. Natl. Acad. Sci.
USA 109, 9705 (2012).10A. D. LaLonde, Y. Pei, and G. J. Snyder, Energy Environ. Sci. 4, 2090
(2011).11Y. Pei, A. D. LaLonde, N. A. Heinz, and G. J. Snyder, Adv. Energy Mater.
2, 670 (2012).12B. A. Volkov, L. I. Ryabova, and D. R. Khokhlov, Phys. Usp. 45, 819
(2002).13D. Khokhlov and B. Volkov, HAIT J. Sci. Eng. 1, 266 (2003).14J. P. Heremans, V. Jovovic, E. S. Toberer, Ali Saramat, K. Kurosaki, A.
Charoenphakdee, S. Yamanaka, and G. J. Snyder, Science 321, 554
(2008).15M. A. Aliyanov, Kh. I. Abilov, and B. B. Kuligv, Phys. Status Solidi A 20,
K17 (1973).16S. Iwanaga, E. S. Toberer, A. LaLonde, and G. J. Snyder, Rev. Sci.
Instrum. 82, 063905 (2011).17Y. Pei, A. D. LaLonde, H. Wang, and G. J. Snyder, Energy Environ. Sci.
5, 7963 (2012).18J. A. Rodriguez-Carvajal, in Satellite Meeting on Powder Diffraction of
the XV IUCr Congress; Vol. 127 (Toulouse, France, 1990).19J. S. Anderson, J. Chem. Soc. 141–143 (1943).20R. F. Bis, J. Phys. Chem. Solids 24, 579 (1963).
FIG. 11. Temperature dependent thermoelectric figure of merit (zT) for the
samples Pb0.999In0.001Te1-ySey (y¼ 0.01, 0.05, 0.10, 0.20, 0.25, and 0.30).
033707-6 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00
21Y. Gelbstein, Z. Dashevsky, and M. P. Dariel, Physicsa B 363, 196–205
(2005).22M. Guch, C. R. Sankar, J. Salvador, G. Meisner, and H. Kleinke, Sci. Adv.
Mater. 3, 615 (2011).23R. I. Ravich, B. A. Efimova, and V. I. Tamarchenko, Phys. Status Solidi B
43, 11 (1971).24Y. I. Ravich, B. A. Efimova, and I. A. Smirnov, Semiconducting Lead
Chalcogenides (Plenum Press, New York, 1970).25Q. Zhang, B. Liao, Y. Lan, K. Lukas, W. Liu, K. Esfarjani, C. Opeil, D.
Broido, G. Chen, and Z. Ren, Proc. Natl. Acad. Sci. 110, 13261 (2013).26J. Androulakis, D.-Y. Chung, X. Su, L. Zhang, C. Uher, T. C. Hassapis, E.
Hatzikraniotis, K. M. Paraskevopoulos, and M. G. Kanatzidis, Phys. Rev.
B 84, 155207 (2011).27D. M. Rowe and C. M. Bhandari, Modern Thermoelectrics (Reston
Publishing, 1983).28W. Kim, J. Zide, A. Gossard, D. Klenov, S. Stemmer, A. Shakouri, and A.
Majumdar, Phys. Rev. Lett. 96, 045901 (2006).
29J. Androulakis, K. F. Hsu, R. Pcionek, H. Kong, C. Uher, J. J. D’Angelo,
A. Downey, T. Hogan, and M. G. Kanatzidis, Adv. Mater. 18, 1170
(2006).30P. F. P. Poudeu, J. D. Angelo, A. D. Downey, J. L. Short, T. P.
Hogan, and M. G. Kanatzidis, Angew. Chem. Int. Ed. 45, 3835
(2006).31A. J. Minnich, M. S. Dresselhaus, Z. F. Ren, and G. Chen, Energy
Environ. Sci. 2, 466 (2009).32C. J. Vineis, A. Shakouri, A. Majumdar, and M. G. Kanatzidis, Adv.
Mater. 22, 3970 (2010).33J.-F. Li, W.-S. Liu, L.-D. Zhao, and M. Zhou, NPG Asia Mater. 2, 152
(2010).34H. J. Goldsmid, Electronic Refrigeration (Pion, 1986), p. 73.35O. Madelung, Semiconductors: Data Handbook (Springer-Verlag, Berlin,
2003).36H. Wang, A. D. LaLonde, Y. Pei, and G. J. Snyder, Adv. Funct. Mater. 23,
1586 (2013).
033707-7 Bali et al. J. Appl. Phys. 116, 033707 (2014)
[This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to ] IP:
131.215.70.231 On: Thu, 11 Sep 2014 14:33:00